Neuron Explorer

Qwen3-0.6B

unclear: prepositions after nouns

A neuron in layer 1 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron weakly responds to small words like 'of', 'in', and 'to' that come right after a noun. The signal is faint and the steering results do not show a clear theme.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. Think it is wrong? Flag it (opens a prefilled issue on GitHub).

Fires most on

The text windows where this neuron's activation was highest. Amber marks positive activation; the outlined token is the peak. Hover a token for its strength.

  1. #1 peak

    َّصَوُّف; صُوفِيّ ṣūfiyy / ṣūfī, مُتَصَوِّف mutaṣawwif), variously defined

  2. #2 peak

    mice into eight separate lines – four lines bred for high levels of daily running, with the

  3. #3 peak

    a living animal that brushed past the resin lost the hair, or the hair was lost in a similar manner by a mammal that came to feed on insects trapped in the resin, which later fossilised into amber.⏎The details of the discovery are

  4. #4 peak

    salmon before settling onto a suitable substrate are now extinct across most of their former range. Highland

  5. #5 peak

    frey, E.Godfrey@stfc.ac.uk⏎Research date: December 2008⏎E Godfrey et al., in Metal 07, proceedings of ICOM-CC metals group conference, Amsterdam (2

  6. #6 peak

    researchers, published in the Journal of Applied Ecology, asserts that poor standardisation of acoustic monitoring limits scientists’ ability to collate data.⏎Kate Jones, chairwoman of the UK-based Bat Conservation Trust⏎told the BBC that “without using the

Steering

Push this neuron up or down while the model writes, and watch what changes. The outputs were generated ahead of time, so nothing runs in your browser.

turned down a lot
Prompt

Here is something I have been thinking about lately: "I am a student who is writing a research paper. I need help with some research paper writing help. I need help with some research paper writing help. I need help with some research paper writing help. I need help with some research

Most negative on

Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.

  1. #1 peak

    with actual tips and suggestions here’s some information to SSL certificates from their early days.⏎It was March 1995 when

  2. #2 peak

    is defined as all activity related to a player’s individual skill development. Based on these definitions, the following is a summary of USA Basketball’s stance on practice/training-to-competition throughout the four levels:⏎

  3. #3 peak

    path which you want to run using Run command. after selecting the path click on the next button.⏎Step 3: Set Run command⏎Now This is important to set run command.⏎Here you have to type the command name which you want to

  4. #4 peak

    worth considering at this time because, sooner or later, similar fights will break out in

Connections

Show on map

Which neurons in earlier layers feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.

Fed by earlier layers

  1. L0:1224 strongly suppresses
  2. L0:178 strongly suppresses
  3. L0:1301 strongly suppresses
  4. L0:1 suppresses
  5. L0:581 weakly excites
  6. L0:2791 weakly suppresses
  7. L0:386 weakly suppresses
  8. L0:83 weakly excites
  9. L0:214 weakly excites
  10. L0:1457 weakly excites
  11. L0:661 weakly suppresses
  12. L0:2195 weakly suppresses
  13. L0:95 weakly excites
  14. L0:29 weakly suppresses
  15. L0:639 weakly excites
  16. L0:621 weakly excites
  17. L0:799 weakly suppresses
  18. L0:915 weakly excites
  19. L0:1992 weakly suppresses
  20. L0:1174 weakly excites
  21. L0:1064 weakly suppresses
  22. L0:615 weakly suppresses
  23. L0:2840 weakly excites
  24. L0:531 weakly excites

Feeds later layers

  1. L15:888 excites
  2. L15:2616 excites
  3. L2:415 weakly suppresses
  4. L2:533 weakly suppresses
  5. L3:522 weakly suppresses
  6. L9:792 weakly excites
  7. L2:1307 weakly suppresses
  8. L9:1943 weakly excites
  9. L11:210 weakly excites
  10. L2:1139 weakly excites
  11. L2:1750 weakly suppresses
  12. L8:985 weakly suppresses
  13. L3:1173 weakly suppresses
  14. L22:1295 weakly excites 'as' before 'well' or 'as much'
  15. L2:1218 weakly suppresses
  16. L16:652 weakly excites
  17. L5:2825 weakly excites
  18. L6:1340 weakly suppresses
  19. L2:1206 weakly suppresses
  20. L3:2511 weakly suppresses
  21. L27:316 weakly suppresses
  22. L17:2413 weakly excites
  23. L5:2758 weakly suppresses
  24. L13:1411 weakly suppresses

Activation histogram

How often this neuron fires at each strength. Almost always it does close to nothing (the tall bars in the middle); the snippets above are the rare moments it fired hard, amber for positive and blue for negative.

2 values between -0.63 and -0.601 values between -0.54 and -0.511 values between -0.51 and -0.481 values between -0.48 and -0.453 values between -0.45 and -0.423 values between -0.42 and -0.397 values between -0.39 and -0.3614 values between -0.36 and -0.3420 values between -0.34 and -0.3118 values between -0.31 and -0.2830 values between -0.28 and -0.2539 values between -0.25 and -0.2252 values between -0.22 and -0.1977 values between -0.19 and -0.1691 values between -0.16 and -0.13132 values between -0.13 and -0.10168 values between -0.10 and -0.07250 values between -0.07 and -0.05392 values between -0.05 and -0.02506 values between -0.02 and 0.01474 values between 0.01 and 0.04374 values between 0.04 and 0.07293 values between 0.07 and 0.10253 values between 0.10 and 0.13161 values between 0.13 and 0.16152 values between 0.16 and 0.19104 values between 0.19 and 0.2288 values between 0.22 and 0.2560 values between 0.25 and 0.2742 values between 0.27 and 0.3030 values between 0.30 and 0.3325 values between 0.33 and 0.3620 values between 0.36 and 0.3911 values between 0.39 and 0.425 values between 0.42 and 0.454 values between 0.45 and 0.481 values between 0.48 and 0.514 values between 0.51 and 0.54offnegativepositiveoften

Logit lens

The words this neuron nudges the model toward (amber) or away from (blue) when it fires. A hint, not a label: many neurons work through later layers instead.

Promotes

  1. venience
  2. ница
  3. ;element
  4. /format
  5. ниц
  6. 啬
  7. �
  8. ETA
  9. iesz
  10. /+

Suppresses

  1. ited
  2. lành
  3. vũ
  4. uffer
  5. apol
  6. idences
  7. pups
  8. erge
  9. liğinde
  10. asymmetric